There are a number of ways to detect high energy radiation. Some of the equipment can be quite bulky, such as a cloud chamber, others may not be as sensitive or quantitative. Scintillator is a very simple but also very accurate method to detect high energy radiation such as x-rays, gamma-rays, high energy particles exceeding a few kilo-electron-volts (KeV) in energy. When high energy radiation strikes on a scintillating crystal, it creates a large number of electron-hole pairs inside the crystal. Recombination of these electron-hole pairs will release energy in the range of a few eV. This energy can be emitted directly from the recombination center as light or transferred to a light emitting ion center which then emits a specific wavelength of light. This low energy emission can then be detected by a photomultiplier tube, avalanche photo diode (APD) or other detector systems with sufficient sensitivity. The higher the light emission (or light yield), the easier for the detector design.
The first scintillating crystal is calcium tungstate (CaWo4) which was used before the turn of this century to detect x-rays. The most significant discovery of a scintillating crystal is Thallium-activated sodium iodide NaI(Tl)) in the mid-40's. Even now, it is still the most widely used scintillating crystal. This is because large size crystals are readily available and quite inexpensive. Moreover, the light yield is the highest among all the known materials and is still the benchmark standard for all other scintillator crystals even after all these years. Even though NaI(Tl)) is widely used, it is not without problems. It is hygroscopic and very soft. Moreover, the density is too low (37 g/cm3), the effective mass number (Zeef) of 49 is also too small. It has a large and persistent after glow which interferes with the intensity counting system. Finally, the decay time about of 230 nanoseconds (ns) is too slow for many applications.
Since the turn of this century; there are a large number of crystals proposed for potential scintillating applications. Even though they do show scintillating properties, none of them has all the right properties. The common problems are low light yield, physical weakness and difficulty in producing large size high quality single crystals. Despite the problems, a number of them have found applications in physics, chemistry, geology and medicine. One common feature of all these usable crystals is that they are the only crystals which can be produced in large size and high quality by an industrial manufacturing process with reasonable cost. This has proven to be the most important factor than the details of scintillating properties to be considered as a viable scintillator material. The specific examples include bismuth germanate ((BGO) which is Bi4Ge3O12), cerium doped gadolinium orthosilicate ((GSO) which is Gd2SiO5) and cerium doped lutetium orthosilicate ((LSO) which is Lu2SiO5).
BGO was found in the early 70's. It has higher density (7.13 g/cm3), and is non-hygroscopic. But it also has problems such as low light yield (15% that of NaI(Tl)), slow decay time (300 ns) and high refractive indices (n=2.15) which results in light loss due to internal reflection. Still BGO scintillator crystals are now used in high energy calorimetry in particle physics research Institutes. It is also used as the detector for the 511 keV gamma-ray radiation of the positron emission tomographs (PET) in medical imaging application. Because of the low light and slow decay, the image produced from the BGO PET machine tends to be blurred with poor resolution.
In early 80's, the Ce doped GSO crystal was disclosed as a scintillator material. It has adequate density (6.71 g/cm3) and is also non-hydrogroscopic. The light yield is 20% of that obtained with NaI(Tl) with a much faster decay time (60 ns). Even though GSO crystals over 80 mm diameter have been produced, the crystal has not yet made it in the PET market because of a strong cleavage plane. It is very difficult to cut and polish the crystal into any specific shape without the risk of fracturing of the entire crystal. Another unexpected problem is the high thermal neutron capture cross-section (49,000 barns) of the gadolinium. It will interfere with the gamma rays generated by neutron irradiation source. However, since there is no neutron source involved in the PET process, gadolinium containing GSO is not a problem.
In the late 80's, the Ce doped LSO crystal was disclosed as a good scintillator material. Similar to GSO, it has high density (7.4 g/cm3) and is non-hygroscopic. The light yield is significantly better and close to 75% that of NaI(Tl)) and the decay time is even faster (42 ns). The index of refraction is also very low (n=1.82). Moreover, since LSO has a totally different crystal structure from GSO, it is fortuitous that in LSO structure, there is not any distinct cleavage plane making the material more suitable for detector block fabrication without the serious risk of fracturing. The thermal neutron capture cross-section is very low (84 barns) as compared to GSO. Lastly, it is now possible to commercially produce high quality, large size single crystals of LSO. Compared with all the other existing known scintillator crystals, Ce doped LSO seems to have the best combination of all the needed properties for PET or other high energy gamma-ray detector application.
Unfortunately, the lutetium element of the crystal contains a trace amount of a natural long decay radioactive isotope, Lu176, which will provide some background count rate that can be harmful for certain highly sensitive detector applications and the crystal has very deep trap centers. This is evidenced by the very long phosphorescence after exposure to any UV light source. The light output measurement of a large number of LSO crystals shows an anti-correlation between trap-related integrated thermoluminescence output and scintillation light output over a range of several orders of magnitude. At present time, the crystal defect is the most serious issue. One can have two crystals with identical appearance with one having 100% light yield and the other failing to scintillate. Thus far, there is no understanding how these deep traps are formed in the first place and there is also no remedy how to reduce or eliminate them.
At present, the scintillation process has been well accepted and used in many applications. The basic mechanism is also reasonably well understood. It is generally accepted that the basic scintillation process involves three steps: (1) the absorption of the incident high energy radiation and the conversion into a large number of low energy (a few multiples of the band gap energy) electrons and hole pairs; (2) transfer the electron-hole recombination energy to the luminescence centers before its loss to multi phonon relaxation processes; and, (3) the radiative emission of the transferred energy. In other word, the scintillation efficiency (L) can be expressed as:E=β×S×Qwhere β is the conversion efficiency, S is the transfer efficiency and Q is quantum efficiency of the radiation centers. Despite the understanding of scintillating mechanism based on the known materials, there is still lack of any good model which has the capability to predict the scintillating behavior of a specific compound. The quantum efficiency of an emission center can be predicted and tested optically; however, neither the total number of electron-hole pairs generated by an incident gamma ray radiation nor the transfer efficiency can be predicted or independently tested. In the end, the only way to confirm the scintillating behavior of a compound is to make and then test it.